Abstract Diffusion MRI (dMRI) allows the in-vivo investigation of the neural architecture of the brain, which can be used to study normal brain development as well as potential pathologies in brain disorders. The spatial resolution of dMRI data sets is around 1.5mm isotropic voxels, which is good to study large and medium size white matter fiber bundles, but grossly insufficient to analyze small fiber pathways. Further, sensitivity to microstructural abnormalities in small cortical and subcortical gray matter structures is lost due to significant partial volume effects that exist at the boundary between different tissue types (e.g., gray-white, gray-CSF, etc.). Thus, a large number of neuropsychiatric disorders cannot be accurately probed at low spatial resolutions. Consequently, we propose several novel acquisition and reconstruction technologies for dMRI that will work synergistically to achieve an order-of-magnitude improvement in dMRI?s spatial resolution, to 600 micron isotropic voxel size. This will provide an extremely detailed in-vivo map of the brain, which will enable new discoveries in white matter connectivity as well as vastly improved sensitivity to small scale tissue abnormalities. This 10-fold improvement in resolution will be achieved in a clinically feasible scan time, on a 3T clinical scanner with high signal quality. The dMRI acquisition development will span i) SNR-efficient acquisition with advanced parallel imaging and specialized RF slab-encoding, ii) navigation-free multi-shot EPI that minimizes geometric distortions and blurring, and iii) motion-robust RF-encoding technique that allow ultra-high resolution dMRI with motion sensitivity exposure time-frame of 2s or less. These technologies will be developed in parallel with a synergistic constrained reconstruction that use phase modeling together with structure-preserving spatial and q- space smoothness constraints, to enable large accelerations while boosting SNR. To ensure scientific rigor, we will comprehensively validate our technology on an ex-vivo human brain along with several healthy volunteers using different quantification metrics. This leap in spatial resolution with acquisition done in a clinically feasible scan time will have a significant and lasting impact in many areas of neuroscience and neurosurgery. For the first time, it will allow accurate and detailed in-vivo investigation of important short cortical association fibers in the superficial white matter regions, as well as functionally critical cortical and sub-cortical gray matter areas. Such technology should also be game-changing to emerging large-scale studies of the brain where dMRI plays a crucial role, such as in the Human Connectome Project, the Adolescent Brain Cognitive Development project, and the U.K. bio-bank project. The ultra-high resolution dMRI will also enhance our ability to understand microstructural abnormalities in neurodevelopmental disorders, and enable accurate delineation of the neural circuitry for positioning the electrode in deep brain stimulation and in image-guided surgery. Thus, we believe that the propose technology will provide a paradigm shift for studying the human brain.